I am using the following code for presenting some measurements, their estimates and their CI:
ssi.oe.cutoffs.chart <- ggplot(data=oe.chart.data, aes(x=core_ot_code_mod, colour=cutoff))
ssi.oe.cutoffs.chart <- ssi.oe.cutoffs.chart + geom_hline(yintercept=1)
ssi.oe.cutoffs.chart <- ssi.oe.cutoffs.chart + geom_errorbar(aes(ymin=lcl95, ymax=ucl95, width=0.5),position="dodge")
ssi.oe.c开发者_运维问答utoffs.chart <- ssi.oe.cutoffs.chart + geom_errorbar(aes(ymin=SIR, ymax=SIR, width=0.5),position="dodge")
core_ot_code_mod
is a discrete variable, and the geom_errorbar
works perfectly for me, showing 2 dodged errorbar with different colour, but I can't use the similar code to let the geom_point
present the same way. And I am using another geom_errorbar
as replacement, which is not very nice.
Can anyone here enlighten me? Thanks!
geom_point requires explicit specification of width of the dodge.
here is a minimal example:
d <- data.frame(expand.grid(x=letters[1:2], co=letters[3:4]), y=runif(4))
d <- transform(d, Lo=y-1, Hi=y+1)
ggplot(d, aes(x, y, colour=co)) +
geom_point(position=position_dodge(width=0.3)) +
geom_errorbar(aes(ymin=Lo, ymax=Hi), width=0.15,
position=position_dodge(width=0.3))
the width=0.15 specifies the width of the errorbar.
the width=0.3 inside position_dodge specifies the width of dodgin.
精彩评论